EP3283343A1

EP3283343A1 - Object tracking before and during an impact

Info

Publication number: EP3283343A1
Application number: EP16710939.6A
Authority: EP
Inventors: Sybille Eisele; Ulf Wilhelm
Original assignee: Robert Bosch GmbH
Current assignee: Robert Bosch GmbH
Priority date: 2015-04-17
Filing date: 2016-03-08
Publication date: 2018-02-21
Anticipated expiration: 2036-03-08
Also published as: CN107531238B; JP6526832B2; US10427677B2; JP2018513053A; EP3283343B1; CN107531238A; WO2016165880A1; US20180043889A1; DE102015207016A1

Abstract

The invention relates to a method comprising steps of detecting an object in an environment of a motor vehicle, tracking the object in relation to the motor vehicle by means of a first motion model of the motor vehicle in a first phase, detecting an impact of the motor vehicle with an obstacle, and tracking the object in relation to the motor vehicle by means of a second motion model of the motor vehicle in a second phase.

Description

description

Object tracking before and during a collision

The invention relates to an environment detection system and a

Crash detection system in a vehicle.

State of the art

A motor vehicle is equipped with a driver assistance system or an automatic control that assists a driver in driving the motor vehicle. In particular, a longitudinal and a lateral control of the motor vehicle are supported individually or jointly. For controlling an environment of the motor vehicle is scanned, for example optically and / or by radar, and objects and the course of their relative movements in the environment of the motor vehicle, so their trajectories are determined from the scanned information. This process is also called tracking. In order to be able to reliably determine the trajectories, the information is usually collected at fixed time intervals and made plausible by means of a movement model. The motion model of the motor vehicle is based on assumptions, such as a maximum acceleration or a maximum yaw rate, which are expected in a normal operation of the motor vehicle.

If the motor vehicle suffers an accident, in particular a collision, then the driver assistance system is usually switched off immediately. The accelerations occurring during a collision frequently exceed the accelerations provided in the movement model by a multiple, so that the plausibility of the trajectories during the accident fails. In order to be able to determine the tracking reliably again after the fall, it is usually necessary to scan the surroundings of the motor vehicle in several scanning steps, whereby a plausibility check usually requires data which has been collected over a period of several seconds or even several minutes. The determination of the trajectories can therefore be poorly or not at all determined even after the end of the accident. The invention has for its object to provide an improved technique for detecting the environment of a motor vehicle in the event of a first collision. The invention achieves this object by means of a method having the features of the independent claim. Subclaims give preferred embodiments again.

Disclosure of the invention

A method comprises steps of detecting an object in an environment of a motor vehicle, tracking, in a first phase, the object relating to the motor vehicle by means of a first movement model of the motor vehicle, detecting a collision of the motor vehicle with an obstacle and the tracking, in a second Phase, of the object with respect to the motor vehicle by means of a second movement model of the motor vehicle.

By applying a different motion model during the collision, detection data of the object may be continuously used to track the object with respect to the motor vehicle. The relative movement between the object and the motor vehicle can thus also be determined in the event of an accident. The movement models usually differ by maximum speed and acceleration values and / or maximum yaw rates of the motor vehicle. For example, the second movement model may allow significantly higher acceleration values than the first movement model.

Preferably, in the second phase, the object is also tracked on the basis of detection data of the object in the first phase. The tracking of the object or its trajectory with respect to the motor vehicle can be made plausible in particular on the basis of detection data of both phases. The object or its trajectory can thus be traced more reliably or completely.

In a further embodiment, the end of the collision is further determined and, in a third phase, the object or its trajectory with respect to the motor vehicle is tracked by means of the first movement model of the motor vehicle. By switching to the usual processing after the At the end of the collision, the detection data of the object can be made more plausible. The object can thus be tracked better or more accurately.

It is particularly preferred that the object in the third phase is also tracked on the basis of detection data of the object in the second phase. In addition or as an alternative to the detection data of the object in the second phase, detection data of the object in the first phase can be used. The aim is the most complete possible use of temporally spaced detection data of the object in several phases. The plausibility of the detection data and thereby the tracking of the object or the course of its relative movement with respect to the motor vehicle can be carried out improved.

It is further preferred that a movement of the motor vehicle in the third phase is controlled on the basis of the tracked object. In particular, the movement of the motor vehicle can be controlled such that the motor vehicle leaves a danger zone that applies in the region of the collision. Other controls can also be performed. For example, it can be determined whether another collision is imminent and measures can be taken to avoid the second collision or to mitigate its consequences. For example, an active or passive safety system can be reactivated for occupants of the motor vehicle.

It can be determined from which direction the obstacle collides with the motor vehicle, wherein the object is tracked in the second phase only on the basis of sensor data whose associated sensors are facing away from the direction of the collision. The second phase is usually too short to perform a plausibility check of sensor or acquisition data. By dispensing with sensor data from such sensors, which are likely to have been impaired by the collision in their functionality, an improved tracking of the object can nevertheless be carried out in the second phase.

In one embodiment, the crash is determined based on data from an acceleration sensor. The accelerometer can be central or be mounted remotely on the motor vehicle. It is also possible to use several acceleration sensors.

In one variant, an acceleration-related acceleration and an acceleration direction are involved in tracking the object in the second phase. The tracking of the object or its trajectory can thus also be carried out under the measuring and processing-technically difficult conditions during the collision.

In a further embodiment, data from an up-front sensor enters the tracking of the object in the second phase. Upfront sensors are usually mounted in front of the vehicle in the direction of travel and can be used, for example, to determine the course and the severity of a frontal impact at an early stage. In addition, a partially covered frontal accident can be detected by means of several Upfront sensors. In the second and third phases, one or more surroundings detection sensors in the area of the affected partial coverage can not be further evaluated when partially overlapped head-on accidents are detected.

Further, data of a peripheral sensor, a roll rate sensor or a yaw rate sensor may be included in the tracking of the object in the second phase. As a result, characteristic data can be used which allow a follow-up of the movement of the object improved.

A computer program product comprises program code means for carrying out the described method when the computer program product runs on a processing device or is stored on a computer-readable data carrier.

Brief description of the figures

The invention will now be described in more detail with reference to the attached figures, in which:

Figure 1 shows a device on board a motor vehicle; FIG. 2 shows phases of a collision of the motor vehicle of FIG. 1 with an object; and

FIG. 3 shows a flow diagram of a method for controlling the motor vehicle of FIG.

Detailed description of embodiments

FIG. 1 shows a motor vehicle 100 with a device 105. The device 105 implements a driver assistance system or an automatic vehicle control, wherein the device 105 may comprise components which are also associated with another system on board the motor vehicle 105, for example a parking or navigation system. In particular, the device 105 is configured to determine an object or the course of its relative movement in an environment with respect to the motor vehicle.

The device 105 comprises a processing device 15, which is equipped with at least one sensor 120 for scanning an object 125 in the environment 110 of the motor vehicle 100. The object 125 may be an object immovable to an environment, such as a street post, or include a moving object, such as another motor vehicle or a pedestrian. Although typically multiple objects 125 are detected, the technique presented below will be described primarily with respect to only one exemplary object 125.

A plurality of sensors 120 can be used which can scan different regions of the environment 110 and / or can be constructed differently. For example, one or more camera or radar sensors 130 may be provided for scanning the object 125. These sensors 120 may be mounted at a central location of the motor vehicle 100 or in the region of an outline of the motor vehicle 100.

For further determination of the movement behavior of the motor vehicle 100, further sensors may be provided. For this purpose, an inertial sensor, which determines an acceleration in one or more spatial directions, or a rotation rate sensor for one or more spatial direction axes belong. For the second phase, crash sensor data can be used to determine the

Crash motion model can be used. For example, a central acceleration sensor 135 and / or one or more upfront sensors 140 and / or one or more peripheral sensors 141, 142 may be used to determine the time of the crash or the crash direction. If at least two upfront sensors 140, for example in the front or in the rear of the motor vehicle 100, are installed in a vehicle, a partial overlap of the motor vehicle 100 with the object 125 during the collision can also be detected. A lateral collision can be detected by means of one or more peripheral sensors and / or a central acceleration sensor. To the peripheral sensors, a pressure sensor 141, which may be installed approximately in a vehicle door, and / or an acceleration sensor, which may be installed, for example, in a Türschweiler or in a B-pillar and / or in a C-pillar, are evaluated. A vehicle rollover can by an additional evaluation of a

Roll rate sensor 143 for determining a rotational rate about the longitudinal axis of the motor vehicle 100 (roll angular rate) are detected. It can also be one

Yaw rate sensor 144 may be provided to determine a rotational speed of the motor vehicle 100 about the vertical axis (yaw rate). The acceleration sensor 135, the roll rate sensor 143 and / or the yaw rate sensor 144 are preferably located on a longitudinal axis of the motor vehicle 100. Several sensors 130-144 can also be designed to be integrated with one another, for example in the form of a multi-channel acceleration sensor.

In one embodiment, the processing device 115 is configured to provide movement information about the object 125 at an interface 150. A driver assistance system or an automatic vehicle control can use data to further control the motor vehicle 100, for example by activating a safety system or also by controlling the movement of the motor vehicle 100. In particular, after the motor vehicle 100 collides with an obstacle 145, the motor vehicle 100 can be driven to leave a danger zone in the environment 110. This danger zone may include the object 125. In another

Embodiment, a safety zone can be actuated which, for example, example, a stanchion or breakdown strip may include a road. For this purpose, a longitudinal or lateral control of the motor vehicle 100 can be actively influenced. Thus, a consequential accident with the motor vehicle 100 can be avoided.

The sensors 120 are typically scanned several times in a predetermined period of time, so that temporally the coordinated measurements are present. These measurements are made plausible with respect to a movement model of the motor vehicle 100. In particular, this movement model may include maximum speed or acceleration values for the motor vehicle 100. Samples indicating, for example, an acceleration of the motor vehicle 100 with respect to the object 125 that exceed the limit of the motion model may be discarded. It is proposed to determine a collision between the motor vehicle 100 and the obstacle 145 and to use an altered movement model during the collision to evaluate the sensor data of the sensors 120. The relative movement of the motor vehicle 100 relative to the object 125 can thus also be determined during the collision correctly and in particular with the aid of sensor values before the collision.

The obstacle 145 may also be considered an object 125 prior to collision, and its movement relative to the motor vehicle 100 tracked. In one embodiment, the tracking of the obstacle 145 as an object 125 may be maintained even after the crash; however, in another embodiment, after the collision, the obstacle 145 is no longer considered and tracked as object 125.

FIG. 2 shows phases of a collision of the motor vehicle 100 of FIG. 1 with the obstacle 145. An exemplary head-on collision is shown, although the technique presented herein is also used with any other type of collision, such as a rear collision, staggered collision, or collision can.

FIG. 2 a shows a first phase in which the motor vehicle 100 is in motion with respect to the obstacle 145 in normal operation. During this first phase, the movement of the motor vehicle 100, as described in more detail above, can be determined by means of a first movement model. FIG. 2b shows the transition from the first to a second phase, during which the motor vehicle 100 collides with the obstacle 145. During the second phase, for example, acceleration values, for example in the longitudinal or transverse direction or high direction or yaw rates about a longitudinal or vertical axis of the motor vehicle 100, may exceed the limits of the first movement model. Such acceleration values or rates of rotation may occur, for example, during a collision, a spin, a rollover or another accident sequence.

Figure 2c shows the motor vehicle 100 in the second phase during the collision with the obstacle 145. In the example chosen, the forward speed of the motor vehicle 100 is rapidly reduced, while the motor vehicle 100 is accelerated sharply around the vertical axis. It is preferred during the second phase to determine the movement of the motor vehicle 100 with respect to a second movement model, which in particular allows acceleration values of the motor vehicle 100, such as may occur during such maneuvers.

FIG. 2 d shows the motor vehicle 100 in a third phase, which can follow the second phase. The collision with the obstacle 145 is finished and the movement determination can be performed again on the basis of the first movement model. In this case, a control of the motor vehicle 100 can be carried out, in particular in order to move the motor vehicle to a safe position or to a safe position. The control can be done by means of a driver assistant or an automatic vehicle control.

FIG. 3 shows a flow diagram of a method 300 for controlling the motor vehicle 100 from FIG. 1. In a first step 305, data of the sensors 20 are recorded at a time interval. The obstacle 145 is tracked in its movement with respect to the motor vehicle 100, wherein a first movement model of the motor vehicle 100 is used. The detected movement may be provided to a driver assistance system or an automatic vehicle control, which in particular performs a longitudinal or lateral control of the motor vehicle 100. The driver assistance system or the automatic The vehicle control system may for example be integrated into the processing device 115 from FIG.

In a step 310 it is determined whether an accident or collision of the motor vehicle 100 with the obstacle 145 has resulted. This determination can be made in particular on the basis of the sensors 135 and / or 140 and / or 141 and / or 142 and / or 143. In a refinement, it may be determined how severe the collision is to deduce from whether the motor vehicle 100 still remains manoeuvrable after the collision has ended. If no collision has been detected, then method 300 may return to step 305 and re-run.

Otherwise, in a step 315, the movement model underlying the tracking of the obstacle 145 with respect to the motor vehicle 100 is changed. The movement model or the movement between the motor vehicle 100 and the obstacle 145 can be determined in particular on the basis of a direction from which the obstacle 145 acts on the motor vehicle 100 and a deceleration of the motor vehicle 100. In addition to the crash acceleration values, yaw rate signals can also be evaluated.

In a step 320, the tracking of the object 125 can be limited to sensor data of such sensors 120, which are remote from the direction of the action of the obstacle 145 on the motor vehicle 100. In practice, there are usually several objects 125, of which then only those can continue to be pursued, which are remote from the direction of action. Thus, it can be avoided that signals from a sensor 140 are used for plausibility or tracking, whose associated sensor 140 has been affected by the collision. In addition, it can thus be avoided that the obstacle 145, which can perform a movement difficult to determine by the collision, is further traced as object 125.

In a step 325, in particular, a position and / or a direction of the motor vehicle 100 after the collision with the obstacle 145 can be determined. The tracking of the object 125 on the basis of the data of the sensors 120 does not have to be made plausible over several measuring cycles for this purpose. In step 330, the end of the crash may be determined, particularly based on acceleration data, such as the acceleration sensor 135 and / or the upfront sensors 140 and / or the peripheral sensors 141, 142 and / or the roll rate sensor 143, with the end of the collision, the first movement model can again be activated in order to further determine the tracking between the motor vehicle 100 and the object 125.

In a step 335, it may be determined whether the motor vehicle 100 is in a danger zone or there is a risk of a following collision. In this case, the motor vehicle 100 may be controlled to assume a safer position in the environment 110. For this purpose, functions of the above-mentioned driver assistance system and / or the automatic vehicle control can be used.

Claims

Method (100), comprising the following steps:

- Detecting an object (125) in an environment (110) of a motor vehicle (100);

Tracking, in a first phase, the object (125) with respect to the motor vehicle (100) by means of a first movement model of the motor vehicle (100);

marked by

- Detecting a collision of the motor vehicle (100) with an obstacle (145);

- Tracking, in a second phase, the object (125) with respect to the motor vehicle (100) by means of a second movement model of the motor vehicle (100).

The method (100) of claim 1, wherein the object (125) in the second phase is also tracked based on detection data of the object (125) in the first phase.

The method (100) of claim 1 or 2, further comprising determining the end of the crash and, in a third phase, tracking the object (125) with respect to the motor vehicle (100) using the first motion model of the motor vehicle (100).

The method (100) of claim 3, wherein the object (125) in the third phase is also tracked based on detection data of the object (125) in the second phase.

Method (100) according to one of the preceding claims, wherein in the third phase a movement of the motor vehicle (100) is controlled on the basis of the tracked object (125).

Method (100) according to one of the preceding claims, wherein it is determined from which direction the obstacle (145) collides with the motor vehicle (100), wherein the object (125) in the second phase is tracked only on the basis of sensor data, their associated sensors of the Facing away from the collision.

The method (100) of any one of the preceding claims, wherein the crash is determined based on data from an acceleration sensor (135).

The method (100) of claim 7, wherein an acceleration due to the collision and an acceleration direction enter into the tracking of the object (125) in the second phase.

The method (100) of any one of the preceding claims, wherein data from an up-front sensor (140) is used to track the object (125) in the second phase.

A method (100) according to any one of the preceding claims, wherein data from a peripheral sensor (141, 142) is included in the tracking of the object (125) in the second phase.

1 1. Method (100) according to one of the preceding claims, wherein data of a roll rate sensor (143) enter into the tracking of the object (125) in the second phase.

The method (100) of any one of the preceding claims, wherein data from a yaw rate sensor (144) is included in the tracking of the object (125) in the second phase.

13. Computer program product with program code means for performing the method (100) according to any one of the preceding claims, when the computer program product runs on a processing device (1 15) or is stored on a computer-readable medium.